SW-Wiktor-Bombola/trees_nn.ipynb

import cv2
import os
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Flatten, Activation, Conv2D, MaxPooling2D
from keras.preprocessing.image import ImageDataGenerator
from sklearn.neural_network import MLPClassifier
from sklearn.model_selection import train_test_split
from tensorflow.keras.optimizers import RMSprop
from sklearn.metrics import classification_report
import re
import matplotlib.pyplot as plt

train_data_dir="../Trees"
validation_data_dir="../Trees"
batch_size=14
img_height, img_width = 60,80

train_datagen = ImageDataGenerator(rescale=1./255,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    validation_split=0.2) # set validation split

train_generator = train_datagen.flow_from_directory(
    train_data_dir,
    target_size=(img_height, img_width),
    batch_size=batch_size,
    class_mode='categorical',
    subset='training') # set as training data

validation_generator = train_datagen.flow_from_directory(
    validation_data_dir, # same directory as training data
    target_size=(img_height, img_width),
    batch_size=batch_size,
    class_mode='categorical',
    subset='validation')

Found 428 images belonging to 3 classes.
Found 105 images belonging to 3 classes.

model = Sequential()

model.add(Conv2D(32, (3,3), activation='relu', input_shape=(60, 80, 3)))
model.add(MaxPooling2D((2,2)))
model.add(Conv2D(64, (3,3), activation='relu'))
model.add(MaxPooling2D((2,2)))
model.add(Conv2D(64, (3,3), activation='relu'))
model.add(MaxPooling2D((2,2)))
model.add(Flatten())
model.add(Dense(512, activation='relu'))
model.add(Dense(512, activation='relu'))

model.add(Dense(3, activation='softmax'))

from tensorflow.keras.utils import plot_model 
!pip install pydot
!pip install graphviz
plot_model(model, to_file='model1_plot.png', show_shapes=True,show_dtype=True, show_layer_names=True, expand_nested=True,)

Defaulting to user installation because normal site-packages is not writeable
Requirement already satisfied: pydot in c:\program files\python39\lib\site-packages (1.4.2)
Requirement already satisfied: pyparsing>=2.1.4 in c:\users\wbloc\appdata\roaming\python\python39\site-packages (from pydot) (2.4.7)
Defaulting to user installation because normal site-packages is not writeable
Requirement already satisfied: graphviz in c:\program files\python39\lib\site-packages (0.19.1)
('You must install pydot (`pip install pydot`) and install graphviz (see instructions at https://graphviz.gitlab.io/download/) ', 'for plot_model/model_to_dot to work.')

print(model.summary())

Model: "sequential_2"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_6 (Conv2D)            (None, 58, 78, 32)        896       
_________________________________________________________________
max_pooling2d_6 (MaxPooling2 (None, 29, 39, 32)        0         
_________________________________________________________________
conv2d_7 (Conv2D)            (None, 27, 37, 64)        18496     
_________________________________________________________________
max_pooling2d_7 (MaxPooling2 (None, 13, 18, 64)        0         
_________________________________________________________________
conv2d_8 (Conv2D)            (None, 11, 16, 64)        36928     
_________________________________________________________________
max_pooling2d_8 (MaxPooling2 (None, 5, 8, 64)          0         
_________________________________________________________________
flatten_2 (Flatten)          (None, 2560)              0         
_________________________________________________________________
dense_6 (Dense)              (None, 512)               1311232   
_________________________________________________________________
dense_7 (Dense)              (None, 512)               262656    
_________________________________________________________________
dense_8 (Dense)              (None, 3)                 1539      
=================================================================
Total params: 1,631,747
Trainable params: 1,631,747
Non-trainable params: 0
_________________________________________________________________
None

model.compile(optimizer=RMSprop(learning_rate=0.0001),
             loss=tf.keras.losses.CategoricalCrossentropy(),
             metrics=['accuracy'])

history = model.fit(train_generator, steps_per_epoch=6, epochs=10,  verbose=1,
                   validation_data = validation_generator, validation_steps = 4)

Epoch 1/10
6/6 [==============================] - 3s 393ms/step - loss: 0.9368 - accuracy: 0.5897 - val_loss: 0.7715 - val_accuracy: 0.6786
Epoch 2/10
6/6 [==============================] - 2s 352ms/step - loss: 0.8682 - accuracy: 0.6310 - val_loss: 0.7431 - val_accuracy: 0.6786
Epoch 3/10
6/6 [==============================] - 2s 352ms/step - loss: 0.7690 - accuracy: 0.6667 - val_loss: 0.6637 - val_accuracy: 0.7143
Epoch 4/10
6/6 [==============================] - 2s 351ms/step - loss: 0.7816 - accuracy: 0.6310 - val_loss: 0.6656 - val_accuracy: 0.6607
Epoch 5/10
6/6 [==============================] - 2s 346ms/step - loss: 0.7070 - accuracy: 0.6667 - val_loss: 0.7131 - val_accuracy: 0.5893
Epoch 6/10
6/6 [==============================] - 2s 351ms/step - loss: 0.6677 - accuracy: 0.6905 - val_loss: 0.5662 - val_accuracy: 0.7143
Epoch 7/10
6/6 [==============================] - 2s 350ms/step - loss: 0.6171 - accuracy: 0.7262 - val_loss: 0.7703 - val_accuracy: 0.5893
Epoch 8/10
6/6 [==============================] - 2s 351ms/step - loss: 0.6001 - accuracy: 0.7738 - val_loss: 0.4881 - val_accuracy: 0.8036
Epoch 9/10
6/6 [==============================] - 2s 362ms/step - loss: 0.4903 - accuracy: 0.7821 - val_loss: 0.4565 - val_accuracy: 0.9107
Epoch 10/10
6/6 [==============================] - 2s 351ms/step - loss: 0.5107 - accuracy: 0.7976 - val_loss: 0.5301 - val_accuracy: 0.8393

plt.plot(history.history['accuracy'], label='accuracy')
plt.plot(history.history['val_accuracy'], label = 'val_accuracy')
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.ylim([0.5, 1])
plt.legend(loc='lower right')

<matplotlib.legend.Legend at 0x1c0ecf5e190>